Radeon HD 3850 and GeForce RTX 4090 Ti - video card comparison
Radeon HD 3850
GeForce RTX 4090 Ti
Architecture
TeraScale (2005−2013)
Ada Lovelace
GPU code name
RV670
AD102
Market segment
Desktop
Desktop
Release date
November 2007
2023
Interface
PCIe 2.0 x16
PCIe 4.0 x16
Supplementary power connectors
1x 6-pin
2x 16-pin
Boost clock speed
-
2565 MHz
Core clock speed
668 MHz
-
Floating-point performance
427.5 gflops
-
Manufacturing process technology
55 nm
5 nm
Number of transistors
666 million
76,300 million
Pipelines / CUDA cores
320
18176
Power consumption (TDP)
75 Watt
600 Watt
Texture fill rate
10.69
1,457
Memory type
GDDR3
GDDR6X
Maximum RAM amount
1 GB
24 GB
Memory bus width
256 Bit
384 Bit
Memory clock speed
1660 MHz
24 GB/s
Memory bandwidth
53.12 GB/s
1,152 GB/s
Display Connectors
2x DVI, 1x S-Video
1x HDMI 2.1, 3x DisplayPort 1.4a
HDMI
-
+
DirectX
10.1 (10_1)
12 Ultimate (12_2)
CUDA
-
8.9
OpenCL
N/A
3.0
OpenGL
3.3
4.6
Shader Model
4.1
6.7
Vulkan
N/A
1.3